The effect of the timing of spring migration on reproductive success differs between the sexes. As a consequence, various sex‐specific tactics relating to the timing of migration have evolved in migratory avian groups. Various hypotheses have been proposed to explain differential migration to breeding or wintering grounds, and inter‐ and intrasexual size differences are often considered one of the proximate mechanisms. We investigated arrival patterns in the spring by individuals of each sex, sexual size dimorphism and related morphological variables, and the relationship between size variation and arrival date in five bunting species that passed through an East Asian migratory flyway stopover site in 2006–08. Males of all the study species arrived before females, and significant sexual dimorphism was observed. Several morphological characters, including total length, wing‐length and tail‐length, contributed to the size variation. Although larger males arrived earlier, there was no relationship between arrival date and size in females. Our study confirmed that East Asian buntings display a discriminated protandrous migration pattern at the stopover site as well as at the breeding grounds. This is consistent with the view that larger body size in males is favoured due to its association with early arrival to help ensure access to the best resources and hence enhanced mating success.

No abstract

No abstract

Abstract. A migrating Black Drongo Dicrurus macrocercus was observed consecutively feeding on a Siberian Stonechat Saxicola maura and an Asian Stubtail Urosphena squameiceps on Hongdo Island, Jeonnam Province, Korea. Unlike previous reports of occasional, apparently exceptional, avian predation by the Black Drongo, this sequential observation suggests that the Black Drongo may selectively hunt avian prey. During Black Drongo migration, when other migrating passerines are abundant and insect availability is relatively low, such behavior would help meet its high energy demands.

No abstract

No abstract

No abstract

Recent AI systems have shown extremely powerful performance, even surpassing human performance, on various tasks such as information retrieval, language generation, and image generation based on large language models (LLMs). At the same time, there are diverse safety risks that can cause the generation of malicious contents by circumventing the alignment in LLMs, which are often referred to as jailbreaking. However, most of the previous works only focused on the text-based jailbreaking in LLMs, and the jailbreaking of the text-to-image (T2I) generation system has been relatively overlooked. In this paper, we first evaluate the safety of the commercial T2I generation systems, such as ChatGPT, Copilot, and Gemini, on copyright infringement with naive prompts. From this empirical study, we find that Copilot and Gemini block only 12% and 17% of the attacks with naive prompts, respectively, while ChatGPT blocks 84% of them. Then, we further propose a stronger automated jailbreaking pipeline for T2I generation systems, which produces prompts that bypass their safety guards. Our automated jailbreaking framework leverages an LLM optimizer to generate prompts to maximize degree of violation from the generated images without any weight updates or gradient computation. Surprisingly, our simple yet effective approach successfully jailbreaks the ChatGPT with 11.0% block rate, making it generate copyrighted contents in 76% of the time. Finally, we explore various defense strategies, such as postgeneration filtering and machine unlearning techniques, but found that they were inadequate, which suggests the necessity of stronger defense mechanisms.Preprint. Under review.